LinkedIn Jobs Scraper - Listings, Multi-Title, Location, CSV

LinkedIn Jobs Scraper is a pay-per-result scraper that turns LinkedIn's public, logged-out job listings into clean, analysis-ready rows — title, company, location, applicant count, advertised pay when the posting shows it, and the full description. It costs $0.001 per job — $1 per 1,000 on the free plan (less on paid plans); you pay for deduped rows only, and empty runs bill nothing. It needs no LinkedIn login, no cookies and no API key: it reads LinkedIn's public guest endpoints directly. Every row carries the same 27 columns, exported as CSV, JSON or Excel.

Try it: Scrape LinkedIn data analyst jobs in London to CSV

✨ What you get

• 🔑 No LinkedIn account, no cookies, no API key — it reads LinkedIn's public guest endpoints directly, and it never auto-applies.
• 📋 27 columns per job, including the full description, applicant count, salary range when the posting shows one, and a 0–100 job score that sorts the freshest, best-documented, least-contested listings first.
• 🎯 Filters applied before you pay — titleInclude / titleExclude, companyNames, excludeAgencies and remote-only are checked on the search cards before the job pages are read, so a listing they drop is neither fetched nor billed.
• 🗺️ Several cities in one run — a locations list (up to 25 places) runs one search per city, the way past LinkedIn's 800-results-per-search wall for a whole state or country; a posting found under two places is delivered and billed once.
• 🔔 Monitoring built in: onlyNewJobs: true delivers and bills only postings you haven't seen; webhookUrl sends a Slack / Discord / webhook digest when new rows land.
• 📤 Exports CSV, JSON or Excel — and runs from the Console, the API, Schedules, n8n / Make / Zapier, or as an MCP tool for AI agents.

It never logs in, so anything LinkedIn keeps behind a sign-in is out of reach: applyUrl is always null, recruiter contact is rare (1 of 40 live job pages measured), applicants is LinkedIn's bucket rather than an exact count and shown on about 55% of postings, and it covers LinkedIn only — Indeed, Glassdoor and company career sites are other actors.

How to scrape LinkedIn jobs in 3 steps

1. Type the job titles and a place — searchQueries (one title or skill per line) and location (a city, state, country or zip; or a locations list to search several cities in one run; or a LinkedIn geoId for an exact pin).
2. Pick the filters you want — date posted, experience level, Easy Apply only, company names or IDs, salary, title words, no agencies — and Start. No LinkedIn login, no cookie, no API key.
3. Download the rows from the Output tab as CSV, JSON or Excel, or call the same input from the API, a Schedule, n8n / Make / Zapier. Set onlyNewJobs: true on a schedule and every later run delivers — and bills — only postings you have not seen.

Key facts:

• Price: $0.001 per job — $1 per 1,000; deduped rows only, empty runs bill nothing.
• No LinkedIn account, no cookies, no API key — it reads LinkedIn's public guest endpoints directly.
• 27 columns per job, including the full description, applicant count, salary range when the posting shows one, and a 0–100 job score.
• Exports CSV, JSON or Excel — and runs from the Console, the API, Schedules, n8n / Make / Zapier, or as an MCP tool for AI agents.
• Monitoring built in: onlyNewJobs: true delivers and bills only postings you haven't seen; webhookUrl sends a Slack / Discord / webhook digest when new rows land.
• Free-plan friendly: Apify's monthly $5 credit covers 5,000 jobs.

Keyword and location in, structured rows out: title, company, location, posting date, employment type, seniority, job function, industry, applicant count, the full job description and the direct job URL. It reads LinkedIn's public guest endpoints directly, with no LinkedIn account, no cookies and no API key.

The minimal input — copy, paste, run:

{ "searchQueries": ["data analyst"], "location": "United States", "maxItems": 100 }

Pricing: $0.001 per job — $1 per 1,000. Deduped rows only; empty runs bill nothing.

Need more than LinkedIn? The same publisher's Multi Job Board Scraper merges LinkedIn + Indeed + Glassdoor + 9 more boards in one run, deduplicated (a job on three boards bills once), with salary annualization, remote-only sweeps across 11 boards, a monitoring mode that only bills postings you haven't seen, and company careers-page watching. This actor is the fast cheap single-board tool; that one is the platform.

Scrape LinkedIn jobs into a clean, analysis-ready spreadsheet. Every row is topped with a 0–100 job score, so the freshest, best-documented, least-contested listings sort first. Built for job seekers, recruiters, job-board builders and labor-market analysts. It reads only public, logged-out listings, and it never auto-applies.

What it does

• Searches LinkedIn's public job listings by keyword and location — read directly from LinkedIn's logged-out search pages (no third-party data provider, no account), with your date posted, Easy Apply and company-id filters applied at the source and experience level checked on each job page (LinkedIn's search stopped applying it on its own, measured 2026-09-25). Give it one location, a locations list (one search per city — the way past LinkedIn's 800-results-per-search wall for a whole country) or a LinkedIn geoId (the numeric place id from a LinkedIn job-search URL, for names like Paris or Cambridge that a text search leaves ambiguous). Employment type and remote/on-site/hybrid are not source filters any more: since LinkedIn's August 2026 AI search (measured 2026-08-29) workplace type and job type are not honoured at the source; the actor converts them to search keywords plus a title/location check — verify with isRemote (and employmentType).
• Filters the result cards before it spends anything on them — titleInclude / titleExclude (title words), companyNames (company name on the card), excludeAgencies (staffing-agency company names) and remote-only are checked on the search cards before the job pages are read, so a listing they drop is neither fetched nor billed.
• Then opens each posting's public job page and fills in what the search result cannot show: employment type, seniority, job function, industry, applicant count, Easy-Apply-or-external, and the whole description as plain text — or as Markdown / HTML with descriptionFormat, rendered from the same page at no extra request. This second pass is what turns a 6-column result list into a full record, and it needs no login either.
• Derives an isRemote flag from remote/WFH wording in the title and location only — e.g. Data Analyst (Remote), United States (Remote). The description is deliberately not scanned: postings that say "there is no option to work remotely", hybrid roles offering "flexibility to work from home when needed", and company boilerplate about "employees working remotely world wide" all read as remote to a text scan, and a wrong remote flag costs you a wasted application. If remote status matters, set remote to ["2"] — the word "remote" is added to the search and only cards whose title or location carries a remote/WFH signal are kept (the same signal isRemote reads), before any job page is fetched. LinkedIn itself no longer filters on it (measured 2026-08-29), so isRemote is never asserted from the filter — it is derived on every row.
• Parses applicant counts to integers so you can filter to low-competition roles with maxApplicants — with the wording LinkedIn actually showed kept alongside in applicantsText.
• Normalizes and cleans: posting dates to YYYY-MM-DD, tracking parameters stripped from job and company URLs, whitespace collapsed, duplicates removed across all your search queries before the job pages are fetched.
• Scores every job 0–100 — freshness up to +28, description depth up to +28, the four job-criteria fields +5 each, low applicant count up to +14, company +5, and +5 for advertised pay — awarded only when the posting's public job page shows a base-pay range (varies by posting and state; measured 2 of 6 pages 2026-08-29), so a posting without advertised pay tops out at 95. Unknowns pay a stated default rather than a zero: unknown posting age scores +8, an unknown applicant count +5, and a posting whose job page could not be read scores +24 for the 48 description-and-criteria points — so a fresh listing whose page was blocked still outranks a month-old one whose page was read, instead of being punished for our fetch luck. The dataset is sorted best-first, ties broken by the fresher posting.
• Applies your row filters after the job pages are read: maxApplicants, requireSalary and minSalary (the last two act on the parsed base-pay range and drop postings that show none — see the FAQ). easyApplyOnly is applied at the source (measured 2026-08-29 from a home IP; the post-fetch check stays as a safety net), so an Easy-Apply run asks LinkedIn for Easy Apply postings instead of reading every job page to find them.

Use cases

• Job search automation — schedule a daily run for your target titles, filter to r86400 (past 24 h) and low applicant counts, and be among the first applicants.
• Recruiting intelligence — see which companies are hiring for a role, at what seniority and employment type, and how much competition each posting draws.
• Job-board aggregation — feed a niche job board with clean, deduped LinkedIn postings including the full description text, ready to render.
• Description mining — descriptionText is the whole posting, so you can grep it for the skills, tools, certifications, visa wording or pay figures you actually care about.
• Labor-market research — track posting volume, seniority mix, industry mix and remote share over time.

Input

{
  "searchQueries": ["data analyst"],
  "location": "United States",
  "datePosted": "r604800",
  "experienceLevel": ["3", "4"],
  "remote": ["2"],
  "maxItems": 100,
  "maxApplicants": 50
}

Output

One dataset row per job, deduped across queries and sorted by job_score. Export to CSV, JSON, or Excel from the Output tab. A real row, unedited apart from the trimmed description and the elided logo URL:

{
  "jobTitle": "Data Analyst",
  "companyName": "Procter & Gamble",
  "location": "Jackson, MS",
  "isRemote": false,
  "employmentType": "Part-time",
  "seniority": "Entry level",
  "jobFunction": "Accounting/Auditing, Administrative, and Analyst",
  "sector": "Government Administration, Education Administration Programs, and Advertising Services",
  "applicants": 25,
  "applicantsText": "Be among the first 25 applicants",
  "applyType": "EASY_APPLY",
  "postedDate": "2026-08-22",
  "postedTimeAgo": "2 hours ago",
  "companyUrl": "https://www.linkedin.com/company/procter-and-gamble",
  "jobUrl": "https://www.linkedin.com/jobs/view/data-analyst-at-procter-gamble-4457816251",
  "descriptionText": "We're looking for a curious, thoughtful Data Analyst who enjoys turning data into answers…",
  "jobId": "4457816251",
  "recruiterName": null,
  "recruiterUrl": null,
  "salaryRaw": null,
  "salaryMin": null,
  "salaryMax": null,
  "salaryPeriod": null,
  "applyUrl": null,
  "job_score": 95,
  "companyLogo": "https://media.licdn.com/dms/image/v2/…/company-logo_100_100/…?e=2147483647&v=beta&t=…",
  "salary_text": ""
}

Which fields does LinkedIn publish, and how often is each one filled?

Every field below is in every row of every run, and the Filled column is measured on live postings rather than estimated, so you know what you are buying before you spend anything.

The Output tab opens on the Overview view (11 columns); the Salary, Apply-ready, Company info and Full record views are one click away — see How to read the output below. Exporting the dataset (CSV / JSON / Excel) always gives you all 27 columns.

If you have an existing integration, two column notes. The posting body ships as descriptionText only: the descriptionHtml column is gone, because it was a second tag-laden copy of the same text and roughly half the size of every row (measured: enriched rows averaged 12.4 KB with it, 5.5 KB without). And the overview view was redrawn on 2026-08-29 — it now shows salary_text (the readable pay) instead of salaryRaw / salaryMin / salaryMax / companyUrl, which moved to the Salary and Company info views — so a pipeline that calls /items?format=csv&view=overview sees the new headers; the raw items never changed shape (salary_text is appended after companyLogo, nothing was renamed or removed).

Two caveats worth knowing before you run it:

• applicants is LinkedIn's bucket, not an exact count: "Over 200 applicants" becomes 200 (a floor), "Be among the first 25 applicants" becomes 25 (a ceiling), and applicantsText keeps the original wording.
• seniority is whatever the employer selected, and most select nothing (34 of 40 live job pages measured), so LinkedIn prints its own placeholder "Not Applicable". That is the source's real answer, so it is passed through rather than blanked; it counts as unset when scoring.

How to read the output

The Output tab has five views of the same rows (nothing is hidden from an export — every row always carries all 27 columns):

• Overview — the first look: company logo, title, company, location, salary, posted date, apply type, applicants, seniority, score, link. Best-scored postings first.
• Salary — the pay columns side by side: the readable Salary, LinkedIn's own wording (Salary (as shown)), and the parsed min / max / period. Filled only when the posting shows a base-pay range.
• Apply-ready — what you need before you apply: Easy Apply vs external, the applicant bucket and its wording, posted date, remote flag, job type, and the job poster when LinkedIn names one.
• Company info — the company side: logo, name, LinkedIn company page, industry, job function.
• Full record — every column in the row's own order, all labelled.

One derived text column, always present, empty ("") when the posting shows no parseable pay — never invented:

• salary_text — built from salaryMin / salaryMax / salaryPeriod, with the currency symbol salaryRaw carries: "$75,000–$150,000 / year", "$45 / hour", "from €60,000 / year", "up to $90,000 / year".

No second date column is added: postedDate is already YYYY-MM-DD (the card's date, null only if the card carries none) and postedTimeAgo keeps LinkedIn's "2 hours ago" wording.

Two derived numbers: job_score (0–100, how it is scored is under What it does; shown in Overview and Full record) and isRemote (from the title and location only; shown in Apply-ready and Full record).

To export just one view: in Console, pick the view in the Output tab, then click Export. From the API, add view=<name> to the items call — view is a documented parameter of Get dataset items — e.g. /v2/datasets/<id>/items?format=csv&view=overview (view names: overview, salary, apply, company, full). Omit view to get all 27 columns.

How much does it cost?

$1 per 1,000 LinkedIn jobs. That is Apify pay-per-event pricing: $0.001 per job on the free plan — paid plans pay less, from $0.0009 on Bronze down to $0.0005 on Diamond — and the live rate is always on this page's Pricing tab. $5 buys 5,000 jobs. On the free Apify plan the monthly $5 credit covers that many jobs, so a real search, or a month of daily onlyNewJobs alerts, costs nothing out of pocket. You are charged only for the cleaned listings delivered after dedup and your filters. Reading each posting's job page costs nothing extra: it adds detail to rows, never rows. No subscription, and no charge for empty runs.

How to get new LinkedIn jobs every day automatically

Schedule this actor (Apify Schedules, daily or every few hours) with onlyNewJobs: true and, if you want to be told rather than poll, a webhookUrl. The first run is the baseline; every later run delivers and bills only postings it has never sent you, and a quiet run costs nothing. A live, copyable setup: New LinkedIn product manager jobs alert (only new postings).

Only new jobs (monitor mode) and partial-run honesty

Set onlyNewJobs: true on a scheduled run and each run delivers only the listings that are new since the previous one — the same search three times a day no longer costs three times. Memory is a named key-value store in your account, one record per (queries, location, filters — plus locations / geoId only when you set them, so an existing schedule keeps its record), keyed on LinkedIn's own jobId, kept 90 days. If the memory cannot be read the run stops and charges nothing rather than re-deliver the whole baseline. If your schedule uses contractType, remote or easyApplyOnly, expect one larger run after 2026-08-29: the first two became search keywords and easyApplyOnly is now applied at the source (LinkedIn's own Easy Apply filter), so postings the old search never surfaced arrive once as new (nothing already delivered is re-sent — the memory key is unchanged).

Every run also writes a RUN_SUMMARY record (key-value store) with delivered, candidates, failed_queries, and a partial map that names any query that stopped early — a mid-pagination HTTP refusal, LinkedIn's 800-result wall, or a sign-in page served instead of listings — so an API caller can tell a short delivery from a small market without reading the status message.

Alerts: get a Slack / Discord / webhook message when new rows land

Set webhookUrl and every run that delivers at least one listing POSTs a digest to it: a Slack incoming webhook gets a text message, a Discord webhook gets a message, any other URL (n8n / Make / Zapier catch hook, your own endpoint) gets JSON {actor, delivered, run_url, dataset_url, rows[:20], text}. Quiet runs send nothing, so pair it with onlyNewJobs: true on a schedule and the actor is an alert service on its own. Delivery is best-effort: a failure is reported in the status message (and RUN_SUMMARY.webhook), never fails the run.

{"searchQueries": ["data analyst"], "location": "London", "datePosted": "r86400", "onlyNewJobs": true, "webhookUrl": "https://hooks.slack.com/services/T000/B000/XXXX"}

Watch a company's LinkedIn postings. companyNames is checked on the search cards before any job page is read, so a scheduled run that combines it with onlyNewJobs and webhookUrl fetches and bills only the watched company's new postings, and messages you only when there are some:

{"searchQueries": ["engineer", "analyst", "manager"], "location": "United States", "companyNames": ["Stripe"], "onlyNewJobs": true, "webhookUrl": "https://hooks.slack.com/services/T000/B000/XXXX"}

companyNames is a substring match on the name LinkedIn prints on the card (Stripe also keeps Stripe Press; a subsidiary under another name is not kept — use companyId for an exact pin), and it is not part of the onlyNewJobs memory key: adding it to an existing schedule narrows what that schedule delivers without re-baselining it. Every card it drops is counted in the status message and in RUN_SUMMARY.card_filters.dropped.company.

Tips / FAQ

How do I scrape LinkedIn jobs without an API?

Give this actor a job title and a place, and it returns LinkedIn's public job listings as rows you can export to CSV, JSON or Excel — with no LinkedIn API key, partner access or account. LinkedIn's own job APIs are partner-gated, so this actor reads the public, logged-out job search and the public job pages instead. The whole input can be three keys: searchQueries, location and maxItems.

Can I scrape LinkedIn jobs without logging in or supplying an li_at cookie?

Yes, this actor needs no LinkedIn account, no li_at cookie and no API key, because it reads only LinkedIn's public logged-out guest endpoints. No credential of any kind is sent to LinkedIn. The only token involved is your Apify token, which starts the run.

Is there a LinkedIn Jobs API?

LinkedIn's own job APIs are partner-gated, and this actor is the API-shaped substitute for the public listings: POST your search as JSON to the run-sync-get-dataset-items endpoint and the HTTP response body is the job rows. Examples in Node, Python and cURL are under API examples below. No LinkedIn key or partnership is involved, just your Apify token.

Why does my run-sync-get-dataset-items call stop at 300 seconds?

Apify's run-sync endpoints cut the HTTP response at 300 s regardless of the run's own timeout. A search reads one job page per second after the listing, so keep maxItems ≤ 150 per sync call, or start the run asynchronously (/runs) and let webhookUrl or an Apify webhook tell you when the dataset is ready.

How much does it cost to scrape 1,000 LinkedIn jobs?

Scraping 1,000 LinkedIn jobs costs $1, which is $0.001 per delivered job. You are charged only for the deduped rows delivered after your filters, and empty runs bill nothing. Reading each posting's job page costs nothing extra: it adds detail to rows, never rows.

Is there a free way to scrape LinkedIn jobs?

Yes — on Apify's free plan the monthly $5 credit covers 5,000 jobs at this actor's rate, so a real search, or a month of daily only-new-jobs alerts, costs nothing out of pocket. No LinkedIn account or API key is needed either.

How do I get only LinkedIn jobs that advertise a salary?

Set requireSalary: true, or set minSalary to a yearly figure (hourly pay is annualized). Both act on the base-pay range parsed from the posting's public job page and drop postings that show none. Pay is published on some postings only (measured 2 of 6 job pages on 2026-08-29), so expect a much smaller run with either switch on.

Why are salaryMin / salaryMax empty on most rows?

Most LinkedIn postings do not advertise pay, so the pay columns are null on every row whose job page shows no base-pay range. When the public job page does show one (measured 2 of 6 pages 2026-08-29), it is parsed into salaryRaw ("$90,000.00/yr - $175,000.00/yr"), salaryMin / salaryMax as numbers, and salaryPeriod (year or hour on every page measured; month, week or day if LinkedIn ever shows them). For postings that state pay only in the body, grep descriptionText.

How do I scrape only remote LinkedIn jobs?

Set remote to ["2"]: the word "remote" is added to the search, and only cards whose title or location carries a remote or work-from-home signal are kept, before any job page is fetched or billed. Verify each row with the isRemote flag. LinkedIn's own workplace-type filter is no longer honoured (see the next answer), so remote status is checked on the card rather than asserted from a filter.

Why don't LinkedIn's remote and job-type filters work any more?

Since LinkedIn's August 2026 AI search, its public job search returns the same results with or without the workplace-type and job-type parameters: measured 2026-08-29 on the keywords "python developer" and location "United States", f_WT and f_JT returned the identical 10 card ids as the unfiltered query. This actor therefore converts remote and contractType into search keywords (remote, hybrid, part-time, contract, temporary, internship) instead of sending parameters the source ignores. Date posted (f_TPR) and Easy Apply (f_AL) still change the result set (re-probed 2026-09-25), so they — and company id (f_C) — are sent. Experience level (f_E) no longer does: on 2026-09-25, f_E = 1, 2 and 4 each returned the same 10 jobs as no filter on "software engineer" and "data analyst" in the United States. experienceLevel is therefore checked on each job page's own "Seniority level" instead (see the input table). Verify the outcome on each row with isRemote and employmentType.

Is applicants an exact number?

No — applicants is the bucket LinkedIn itself shows, and it shows one on about 55% of postings. applicantsText keeps the exact wording ("Over 200 applicants", "Be among the first 25 applicants") next to the integer.

Why does seniority say "Not Applicable" so often?

"Not Applicable" is LinkedIn's own text for "the employer didn't pick one", and most don't: 34 of 40 live job pages measured carried no employer-set seniority. It is passed through unchanged rather than blanked so you can tell "unset" apart from "we failed to read it", and it counts as unset when scoring.

What's the difference between jobFunction, sector and isRemote?

jobFunction is the role area LinkedIn lists (for example "Information Technology") and sector is the employer's industry, and both come straight from the job page. Remote status is not a public field, so isRemote is derived from wording in the title and location only, never from the description, which is full of hybrid arrangements, work-from-home perks and company boilerplate that a text scan reads as "remote". Treat isRemote as a hint, unless you set remote: ["2"], in which case every delivered row was checked against that title and location signal before its job page was fetched.

Can I get the description as Markdown or HTML?

Yes, set descriptionFormat to "markdown" or "html" and the posting body ships in that format, in the same descriptionText column and the same position. The job page is fetched anyway, so the format is rendered from the same response: no extra request, no price change. Markdown keeps bold, italics, bullets, numbered lists, links and line breaks from LinkedIn's markup, which is what an LLM or a spreadsheet reads cleanly; bold or italic that LinkedIn stretches across a paragraph break is closed and reopened per line, so every ** pair renders, and nested lists and tables are flattened to text. HTML is the block as LinkedIn serves it, with only its Angular comment markers removed. job_score and isRemote are identical across all three formats (measured 2026-08-29 on a 10-row London run: 10 of 10 Markdown rows clean, no tag or entity left).

How many LinkedIn jobs can I scrape in one run?

LinkedIn's public search serves at most 800 results per search. To cover a whole state or country, pass a locations list of up to 25 places: one search per city, and a posting found under two places is delivered and billed once. maxItems (1 to 1000) applies to each query and, with locations, to each place.

How fresh are the results?

Every run reads LinkedIn's live public search at the moment it starts, so the rows are the listings LinkedIn is showing right then, not a cached index or a resold dataset. Narrow to recent postings with datePosted: r86400 (past 24 h), r604800 (past week) or r2592000 (past month). Each row carries the card's own date as postedDate (YYYY-MM-DD, null only when the card shows none) plus LinkedIn's postedTimeAgo wording.

Why does a LinkedIn scraping run take a minute or two per hundred jobs?

After the search it reads each posting's own job page, roughly one per second, because that page is the only public source for seniority, employment type, job function, industry, applicants and the description. LinkedIn throttles on concurrency rather than volume: measured on 40 postings, one request at a time enriched 40 of 40 (100%) at 0.96 rows per second, while two at a time enriched only 32 of 40 (80%) at 0.73 rows per second. So it goes one at a time on purpose. If LinkedIn blocks the job pages anyway, rows still ship with the search-card fields and the run still succeeds, and the status message tells you how many full records you got and which of the possible reasons actually applied (rate-limited, served a sign-in page, returned 4xx, or never requested because the run ran out of time), so you know whether retrying will help.

How do I turn LinkedIn job search into a daily alert of only new postings?

Schedule this actor with onlyNewJobs: true and a webhookUrl, and every run after the first delivers, bills and messages you only about postings it has never sent you. The first run is the baseline, and a run with nothing new costs nothing and sends nothing. A live, copyable setup: New LinkedIn product manager jobs alert (only new postings). Details in How to get new LinkedIn jobs every day automatically above. The same mode across LinkedIn plus 11 more boards is the Multi Job Board Scraper.

Can I run it from n8n, Make, Zapier or an AI agent?

Yes, every input works identically from the API, so the actor drops into n8n, Make, Zapier, an Apify Schedule or your own pipeline unchanged. In n8n, use the Apify node with actor flash_scraper/linkedin-jobs-scraper, or an HTTP Request node on the endpoint below. In Make and Zapier, use the Apify app's "Run an Actor" module. AI agents can call it as a tool through Apify's MCP server. The input schema above is the whole interface, and nothing is Console-only.

How do I export LinkedIn jobs to CSV or Excel?

Run the actor, then download the rows from the Output tab as CSV, JSON or Excel; every export carries all 27 columns. The same rows come out of the Dataset API for pipelines, and adding view=overview (or salary, apply, company, full) exports one named subset of the columns.

Does it scrape Indeed or Glassdoor as well?

No, this actor reads LinkedIn only. For one search across LinkedIn, Indeed, Glassdoor and 9 more boards, deduplicated so a job found on three boards bills once, use the same publisher's Multi Job Board Scraper.

What can't this actor get from LinkedIn?

It never logs in, so anything LinkedIn keeps behind a sign-in is out of reach. In detail:

• applyUrl is always null: an external posting's Apply button is a sign-in prompt with no destination link in the public markup.
• Recruiter contact (recruiterName, recruiterUrl) is rare — 1 of 40 live job pages measured — and only filled when the poster enabled "message the job poster".
• applicants is a bucket rather than a count, and LinkedIn shows it on about 55% of postings.
• Workplace type and job type cannot be filtered at the source any more, so they are handled as search keywords and verified on the row.
• It covers LinkedIn only. Indeed, Glassdoor and company career sites are other actors.

Does it apply to jobs for me?

No, it never submits an application, including Easy Apply. It only collects public listings into a dataset, and it never logs in.

Where does the data come from, and what if the source is down?

Rows come straight from LinkedIn's public, logged-out job search and job pages, with no intermediary and no resold dataset. If one query fails it is skipped and the run continues with the rest. If every request fails or is refused, the run ends successfully with a status message that says so, tells you to retry in a few minutes, and confirms you were not charged.

Is scraping LinkedIn jobs legal?

This actor collects public job data only, from logged-out pages, and it never signs in or bypasses a gate. LinkedIn rate-limits aggressively and restricts automated access in its terms, so expect occasional partial runs. Use the data for personal, research or recruiting purposes, and follow local law.

Related actors

• Workday Jobs Scraper — live careers from any company hiring through Workday, no API key
• ATS Job Scraper — Greenhouse, Lever & Ashby company boards with only-new-jobs alerts
• Multi-Jobboard Scraper — one search across several job boards at once
• Remote Job Aggregator — RemoteOK, WeWorkRemotely + 8 more remote boards in one deduplicated feed, $3 per 1,000
• Company & Domain Enricher — turn hiring companies into full firmographic records

Support: found a bug or need a feature? Open an Issue on this actor's Issues tab — typical response within 1 business day.

API examples

Run it from code exactly like the Console — same input keys.

// Node.js (apify-client)
const { ApifyClient } = require('apify-client');
const client = new ApifyClient({ token: 'YOUR_APIFY_TOKEN' });
const run = await client.actor('flash_scraper/linkedin-jobs-scraper').call({
    searchQueries: ['data analyst'], location: 'Austin, TX', maxItems: 100,
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items.length, 'jobs');

# Python (apify-client)
from apify_client import ApifyClient
client = ApifyClient("YOUR_APIFY_TOKEN")
run = client.actor("flash_scraper/linkedin-jobs-scraper").call(run_input={
    "searchQueries": ["data analyst"], "location": "Austin, TX", "maxItems": 100,
})
items = client.dataset(run["defaultDatasetId"]).iterate_items()

# cURL
curl -X POST "https://api.apify.com/v2/acts/flash_scraper~linkedin-jobs-scraper/run-sync-get-dataset-items?token=YOUR_APIFY_TOKEN" \
  -H 'Content-Type: application/json' \
  -d '{"searchQueries":["data analyst"],"location":"Austin, TX","maxItems":100}'

Use it from an AI agent (MCP)

Apify's MCP server exposes this actor as a tool named after its slug, flash_scraper/linkedin-jobs-scraper, with the slash replaced — the server's README lists Actor tools as flash_scraper--linkedin-jobs-scraper (its apify--rag-web-browser form); the Apify docs page also writes the older flash_scraper-slash-linkedin-jobs-scraper form, so look for either in your client's tool list. Point any MCP client (Claude, Cursor, VS Code, Codex…) at https://mcp.apify.com?tools=flash_scraper/linkedin-jobs-scraper to load just this tool. A minimal call:

{"searchQueries": ["data analyst"], "location": "London", "maxItems": 10}

Measured 2026-08-29 from a home IP: 10 rows with the full record in 9 seconds. Price: $0.001 per delivered job (the live free-plan pay-per-event rate on this page's Pricing tab; paid plans get a small tiered discount), so that call costs $0.01. With Apify's x402 agentic payments an agent pays in USDC on Base for a prepaid token (minimum $1) that each run draws from, and needs no Apify account ("No Apify account is required" — Apify's own words).

Schedules & pipelines

Every input works identically from the API, so it drops straight into an Apify Schedule (daily exports), webhooks (push finished runs to Slack or Sheets via Zapier/Make/n8n), or your own pipeline via the Dataset API. For a scheduled feed that only ever bills NEW postings, set onlyNewJobs: true — see How to get new LinkedIn jobs every day automatically.

Proxy (optional)

No proxy is needed: by default every request goes straight from the Apify container, and LinkedIn's public job search answers that today. Leave proxyConfiguration empty and nothing changes.

Reach for a proxy when LinkedIn gates your region (runs that end with "answered with a sign-in page" or HTTP 999/403 on every request) or on heavy schedules (many queries at maxItems: 1000, several times a day), where one exit IP gets throttled. Pass the standard Apify proxy object —

{"useApifyProxy": true, "apifyProxyGroups": ["RESIDENTIAL"], "apifyProxyCountry": "US"}

Every run comes with a report

Every run that delivers at least one listing also writes a REPORT record to its key-value store: one self-contained HTML page with the headline numbers (listings delivered, share with advertised pay, companies, Easy Apply share, full-record share, remote share when at least one row is remote, and new-since-last-run on an onlyNewJobs schedule), bar charts of the top companies, the seniority split and Easy Apply vs external, and the top rows in the Overview view's columns with the same headers (Title, Company, Location, Salary, Posted, Apply type, Applicants, Seniority, Score, Link; the logo is left out of the static table, and so is any column that is empty on every shown row — on a run where no posting advertises pay there is no Salary column). Open it from the run's Output tab → REPORT, or follow the Report: link at the end of the run's status message (RUN_SUMMARY.report_url for API callers). No scripts, no external assets — safe to forward or screenshot as-is. The dataset stays the source of truth: the report shows at most 100 rows and links back to the full dataset.

Related Flash Scrape actors

• Multi Job Board Scraper — LinkedIn, Indeed, Glassdoor + 9 more boards, deduplicated, with only-new job alerts
• Local Business Leads Scraper — any category, any city, verified emails, lead scores
• Remote Jobs Aggregator — 10 keyless remote boards in one deduplicated feed
• Creator Leads Scraper — TikTok, Instagram and YouTube creator emails in one run
• Shopify Store Scraper — any store's full public catalog, every variant, SKU and price, plus a store-intelligence row
• All Flash Scrape actors — same house rules everywhere: pay per delivered row, honest status messages, only-new monitoring, webhook alerts, and a run report on every run.